GC-analysis 0.4.5

A program that compute the GC percentage of a given genomic sequence

GC-analysis

A command-line utility for calculating GC percentages of genome sequences

Quick starter

Calculate the GC content of chromosome 17 of the human reference genome with window size (or span) = 5 and shift (or step) = 5. Input fasta file is GRCh38-Chrom17.fasta and output wiggle file is GRCh38-Chrom17.wig. Note that the output file's extension is added by the program.

~ $ GC_analysis -i GRCh38-Chrom17.fasta -w 5 -s 5 -o GRCh38-Chrom17

Installation guide

Note that pyBigWig can only be used under linux environment. To work with Windows system, the Docker image can be used as shown below. Alternatively, you can clone the repository, comment out import pyBigWig and the script would work but without BigWig support.

1. Pip install GC_analysis (NB. Python3 is recommanded but GC_analysis should work with Python2 as well)

pip3 install GC_analysis

Then GC_analysis.py command will be available globally.

GC_analysis.py -i [INPUT] -o [OUTPUT] -w [window size] -s [shift]

2. Run the python script directly. Please ensure you have python3 installed with pyBigwig and Biopython. Clone the github repository and install packages.

git clone https://github.com/tonyyzy/GC_analysis
cd GC_analysis
pip3 install -r requirements.txt

run the script from GC_analysis directory.

python3 ./GC_analysis/GC_analysis.py -i [INPUT] -o [OUTPUT] -w [window size] -s [shift]

3. Use the packaged binary.

mkdir ~/GC_analysis
cd ~/GC_analysis
wget https://github.com/tonyyzy/GC_analysis/releases/download/v0.3/GC_analysis

Execute the binary command

GC_analysis -i [INPUT] -o [OUTPUT] -w [window size] -s [shift]

4. Use the Docker image. Firstly, pull the docker image (around 384 MB)

docker pull tonyyzy/gc_analysis

To use input files outside the container and save output files on your computer, the -v volume mapping option will be used. You will need to know the absolute path of the directory you want to map (which can be found out with pwd).

docker run -v /your/local/path:/app tonyyzy/gc_analysis GC_analysis -i /app/yours.fasta -o /app/yours -w 5 -s 5

This option maps /your/local/path to /app under the container's root directory. Your result file will be saved to /your/local/path/yours.wig.

Command-line options

~ $ GC_analysis -h
usage: GC_analysis [-h] -i INPUT_FILE -w WINDOW_SIZE -s SHIFT [-o OUTPUT_FILE]
                   [-ot] [-f {wiggle,gzip,bigwig}]

required named arguments:

-i INPUT_FILE, --input_file INPUT_FILE
INPUTFILE: Name of the input file in FASTA format

-w WINDOW_SIZE, --window_size WINDOW_SIZE
WINDOW_SIZE: Number of base pairs that the GC percentage is calculated for

-s SHIFT, --shift SHIFT
SHIFT: The shift increment (step size)

optional arguments:

-h, --help
Show the help message and exit

-o OUTPUT_FILE, --output_file OUTPUT_FILE
OUTPUT_FILE: Name of the output file

-ot, --omit_tail
Use if the trailing sequence should be omitted. Default behaviour is to retain the leftover sequence.

-f {wiggle,bigwig,gzip}, --output_format {wiggle,bigwig,gzip}
Choose output formats from wiggle, bigwig or gzip compressed wiggle file.

Example usage

1. Calculate the GC content of chromosome 17 of the human reference genome, the percentage is calculated over five base pairs (window_size), and the window is shifted by five base pairs every time (i.e. there is no overlapping base paires in each entry).

~ $ GC_analysis -i GRCh38-Chrom17.fasta -w 5 -s 5 -o GRCh38-Chrom17

2. By default, the GC percentage of the trailing sequence is calculated and appended to the end of the output file. For example, with the following input

~ $ GC_analysis -i examaple1.fasta -w 5 -s 5 -o with_tail

and example1.fasta is

>chr1
AAAAACC

the generated with_tail.wig will look like

track type=wiggle_0 name="GC percentage" description="chr1"
variableStep chrom=chr1 span=5
1	0
6	100

If it is desirable to omit the trailing sequence in the result, the -ot or --omit_tail option can be used. For example

~ $ GC_analysis -i examaple1.fasta -w 5 -s 5 -o without_tail -ot

will generate output file without_tail with the following content

track type=wiggle_0 name="GC percentage" description="chr1"
variableStep chrom=chr1 span=5
1	0

3. The program support three output file formats, wiggle, bigwig and gzip compressed wiggle file. Wiggle output file follows the UCSC variableStep format definition. Wiggle file is the default output format. The output format can be changed with -f or --format option.

~ $ GC_analysis -i GRCh38-Chrom17.fasta -w 5 -s 5 -o GRCh38-Chrom17

and

~ $ GC_analysis -i GRCh38-Chrom17.fasta -w 5 -s 5 -o GRCh38-Chrom17 -f wiggle

will generate GRCh38-Chrom17.wig as the output file.

~ $ GC_analysis -i GRCh38-Chrom17.fasta -w 5 -s 5 -o GRCh38-Chrom17 -f gzip

will generate GRCh38-Chrom17.wig.gz as the output file. Decompress GRCh38-Chrom17.wig.gz will give you the same wiggle file as choosing wiggle as the output format.

~ $ GC_analysis -i GRCh38-Chrom17.fasta -w 5 -s 5 -o GRCh38-Chrom17 -f bigwig

will generate GRCh38-Chrom17.bw as the output file. It should be noted that bigwig format does not allow overlapping bases, which means that -w 5 -s 3 is an invalid option with choosing bigwig as the output format. In this case, where shift is smaller than window size and bigwig format is specified, the program will generate a wiggle file instead and output a warning message.

~ $ GC_analysis -i GRCh38-Chrom17.fasta -w 5 -s 3 -o GRCh38-Chrom17 -f bigwig
WARNING! BigWig file does not allow overlapped items. A wiggle file was generated instead.

4. If an output filename is not given, the result will be written to stdout. If the output filename is not given and a file format other than wiggle was chosen, the program will automatically output the result to stdout and give you a warning before and after the result. Eg.

GC_analysis -i example1.fasta -w 5 -s 3 -f bigwig
WARNING! BigWig file does not allow overlapped items. A wiggle file will be generated instead.
WARNING! An output filename is needed to save output as bigwig. The result is shown below:
track type=wiggle_0 name="GC percentage" description="chr1"
variableStep chrom=chr1 span=5
1       0
4       50
WARNING! BigWig file does not allow overlapped items. A wiggle file was generated instead.
WARNING! An output filename is needed to save output as bigwig. The result is shown above.

Timing againts human chromosomes

Click for raw data table

Entry	Human chromosome	No. of base pairs	Average real time - single thread (s)	Average real time - multi threads (s)
CM000663.2.fasta	1	248956422	288.429	179.221
CM000664.2.fasta	2	242193529	276.355	169.611
CM000665.2.fasta	3	198295559	227.528	135.637
CM000666.2.fasta	4	190214555	217.846	153.091
CM000667.2.fasta	5	181538259	205.623	123.858
CM000668.2.fasta	6	170805979	193.209	117.180
CM000669.2.fasta	7	159345973	183.445	109.135
CM000670.2.fasta	8	145138636	166.607	98.632
CM000671.2.fasta	9	138394717	157.142	93.898
CM000672.2.fasta	10	133797422	150.872	92.371
CM000673.2.fasta	11	135086622	154.003	92.498
CM000674.2.fasta	12	133275309	150.533	90.807
CM000675.2.fasta	13	114364328	129.951	77.498
CM000676.2.fasta	14	107043718	121.008	71.970
CM000677.2.fasta	15	101991189	115.194	68.336
CM000678.2.fasta	16	90338345	103.169	60.799
CM000679.2.fasta	17	83257441	94.353	55.729
CM000680.2.fasta	18	80373285	92.020	53.395
CM000681.2.fasta	19	58617616	67.506	39.308
CM000682.2.fasta	20	64444167	74.048	43.280
CM000683.2.fasta	21	46709983	53.633	31.118
CM000684.2.fasta	22	50818468	57.466	33.701
CM000685.2.fasta	X	156040895	176.895	105.408
CM000686.2.fasta	Y	57227415	67.016	38.142
J01415.2.fasta	MT	16569	0.231	0.397

* 1) Real time data is the average of three runs; 2) GC_analysis parameters for each run is -w 5 -s 5; 3) Serial data is collected with the Master branch, Parallel data is collected with the Parallel branch.

As can be seen from the plot, GC_analysis scales well with number of base pairs, resulted a linear relationship between the execution time and the size of the chromosomes. Although multi-threaded version can provide ~1.7x speed improvement, it has a significantly higher memory consumption, hence it's not recommended.

(EXPERIMENTAL!!!) Multi-threaded GC_analysis

Git clone the parallel branch from GitHub repo:

git clone --single-branch -b parallel https://github.com/tonyyzy/GC_analysis

Execute as normal from GC_analysis directory

~ python3 ./scripts/GC_analysis.py -i GRCh38-Chrom17.fasta -w 5 -s 5 -o GRCh38-Chrom17

This multithreading implementation is a very crude one and only result in ~1.7x speed up. A large amount of RAM is needed to store out-of-order intermediate results for sorting.